Neural Network / Biological & Artificial Neural Network / BNN & ANN
Hossein Banki-Koshki; Seyyed Ali Seyyedsalehi
Volume 15, Issue 3 , December 2021, , Pages 199-209
Abstract
The presentation of new neuronal models to simulate cognitive phenomena in the brain has attracted the research interests in recent years. In this study, a new neural model based on the chaotic behavior of weights of artificial neural networks during training by back-propagation algorithm is presented. ...
Read More
The presentation of new neuronal models to simulate cognitive phenomena in the brain has attracted the research interests in recent years. In this study, a new neural model based on the chaotic behavior of weights of artificial neural networks during training by back-propagation algorithm is presented. This model is the first discrete neuronal model with learning ability and shows complex and chaotic behaviors. The learning ability of this model has enabled it to simulate cognitive phenomena such as neuronal synchronization in near-realistic conditions. The model, which is derived from a simple three-layered feed-forward neural network, has several coexisting attractors that make learning possible in various basins of attraction. The study of model parameters shows that bifurcation occurs not only by changing the learning rate, but also external stimulation can change the model behavior and bifurcation pattern. This point that can be used in modeling and designing new therapies for cognitive disorders.
Neural Network / Biological & Artificial Neural Network / BNN & ANN
Seyedeh Sadaf Razavinezhad; Amir mohammad Fallah; Seyed Abolghasem Mirroshandel
Volume 14, Issue 4 , February 2021, , Pages 307-320
Abstract
Diabetes is a common disease all around the world. It is a difficult and incurable but controllable disease, so it is important to control and prevent its complications. Thus, low error and smart methods are used to predict blood glucose levels and prevent dangerous complications to control it effectively. ...
Read More
Diabetes is a common disease all around the world. It is a difficult and incurable but controllable disease, so it is important to control and prevent its complications. Thus, low error and smart methods are used to predict blood glucose levels and prevent dangerous complications to control it effectively. In this regard, different methods were investigated. In this research, two models based on deep learning technique are used which produce efficient and optimal results. These models are composed of different combinations of long short-term memory and feed forward neural networks which predict the patient's future blood glucose levels with considerable accuracy and speed. To achieve more comprehensive model, 81,200 blood glucose data was evaluated through 203 patients. In addition, 27 effective features in patients' blood glucose levels are considered in this regard. Furthermore, cross-validation method which is suitable for time series was used for more accurate evaluation. The results showed that Autoregressive Integrated Moving Average model could not predict blood glucose levels considering this amount of data and system hardware limitations, while the models based on deep learning had good performance and good speed. Furthermore, the second proposed model for the prediction horizons of 5, 10, and 15 minutes outperformed the first one with 13.8%, 16%, and 18.9%, respectively. Therefore, the second proposed model can be more reliable for predicting blood glucose. So, it can be used in smart warning systems to prevent hypoglycemia, which is a dangerous and widespread problem of type 1 diabetes.
Neural Network / Biological & Artificial Neural Network / BNN & ANN
Hamed Abbasi; Shahrokh Shojaei; Nasim Naderi
Volume 13, Issue 2 , August 2019, , Pages 105-115
Abstract
Today, in order to decide on many cardiac surgeries, and whether the patient is able to get under surgery or the time of surgery is passed, it is necessary to measure pulmonary vascular resistance and if the resistance is above a threshold, the patient is considered to be non-surgery; and sometimes, ...
Read More
Today, in order to decide on many cardiac surgeries, and whether the patient is able to get under surgery or the time of surgery is passed, it is necessary to measure pulmonary vascular resistance and if the resistance is above a threshold, the patient is considered to be non-surgery; and sometimes, some therapies are used to reduce the resistance of the pulmonary arteries to the initial disease of the arteries, in which, in order to track down the resistance of the pulmonary vascular, a re-measurement of this parameter is required. Currently, the golden standard of this measure is the use of catheterization procedures, which are aggressive and associated with complications. The purpose of this study is to replace a non-invasive method, rather than an invasive method of cardiac catheterization, by predicting pulmonary vascular resistance based on echocardiographic data by artificial neural networks. Research was performed on 591 patients. Echocardiography was recorded for all subjects, and the echocardiographic data (mPAP, dPAP, sPAP, PCWP, CO) as the neural network input and pulmonary vascular resistance of all patients who were subjected to previous catheterization was evaluated as the output of the neural network and thus, it was obtained, the relationship between echocardiography data and PVRcath. The proposed neural network was typically learned with 75% of the data, and was tested with 25% of the data, and these ratios were modified to better learn the neural network. As a result of implementation, the mean squared error, respectively, for the learning and testing data for the proposed neural network, was 0.37 and 0.27 for the first model, 14.67 and 10.76 for the second model, and 15.82 and 9.58 for the third model.